Oppgrader til nyeste versjon av Internet eksplorer for best mulig visning av siden. Klikk her for for å skjule denne meldingen
Ikke pålogget
{{session.user.firstName}} {{session.user.lastName}}
Du har tilgang til Idunn gjennom , & {{sessionPartyGroup.name}}

4. Antisemitism and Islamophobia in Norway

A Survey Analysis of Prevalence, Trends and Possible Causes of Negative Attitudes towards Jews and Muslims

Ottar Hellevik (b. 1943) is Professor Emeritus in Political Science at the University of Oslo, where he has worked since graduating in sociology in 1968. He has been Chairman of the department and Dean of the Faculty of Social Sciences. Fields of research include survey methodology and analysis; values and value change; attitudes and attitude change; social inequality; and political recruitment. Among his publications are the textbooks Forskningsmetode i sosiologi og statsvitenskap (Research Methods in Sociology and Political Science) (Universitetsforlaget 1971, 7th ed. 2002) and Introduction to Causal Analysis (George Allen & Unwin 1984, Scandinavian University Press, 1988).

The aim of the chapter is to establish how widespread negative attitudes towards Jews and Muslims are among the Norwegian population, and to look for factors that may stimulate such attitudes, through an analysis of the two representative population surveys conducted by The Norwegian Center for Holocaust and Minority Studies in 2011 and 2017.

Attitudes towards Jews are measured by indices of prejudice, dislike, social distance, and a summary index of antisemitism. Islamophobia is measured by a corresponding set of indices in 2017. The level of negative attitudes towards Jews is low and declining for all indices. Negative attitudes towards Muslims are more widespread. Women, younger people and those with higher education have a lower level of negative attitudes towards the two minorities. Opinion on the Middle East conflict affects antisemitism and Islamophobia in opposite directions, while both are strongly influenced by xenophobia. Negative attitudes towards the two minorities tend to coexist in individuals.

Keywords: Islamophobia, antisemitism, population survey, xenophobia, attitudes towards the Israeli-Palestinian conflict, Norway

1. Outline of chapter

The analyses use data from the two representative population surveys from 2011 and 2017, conducted by The Norwegian Center for Holocaust and Minority Studies (CHM). Each survey had a little over 1,500 respondents (section 2). Indices of prejudice, dislike and social distance, which are combined in a summary index of antisemitism, were measured in the same way in both surveys (section 3). For Islamophobia, a corresponding set of indices was used in 2017 (section 4).

The importance of the respondents’ own attitudes for how they perceive the prevalence of negative attitudes towards the two minorities in Norway, and whether they see a need to combat such attitudes, is analysed in section 5. The relationship between attitudes towards the two minorities, whether they are opposing or go together at the individual level, is the topic of section 6.

Variations in attitudes towards Jews and Muslims depending on respondent characteristics such as social background, religiosity, opinion regarding the conflict between Israel and the Palestinians, xenophobia and scepticism towards immigrants are studied in order to shed light on possible causes for antisemitism and Islamophobia (section 7). Changes in these variables and their contribution to the effect of generational replacement and individual changes in attitudes on the trend for antisemitism between 2011 and 2017 are analysed (section 8). The concluding section (9) summarises the main findings from the analyses.

2. The data

The population surveys were conducted electronically using GallupPanelet, Kantar TNS’s access panel. The sample members received email invitations to complete a web questionnaire.1 The gross samples were stratified prior to distribution and selected in proportion to the Norwegian population’s distribution by education, gender, age and geographical region. Weights were calculated to correct for observed biases with regard to these variables in the net sample.2

One reminder was issued during the field period. The number of interviews obtained was 1,522 in 2011 and 1,575 in 2017 (response rates 48% and 54%).3 The interviews took place in November 2011 and January to April 2017. In 2017, three minority samples (Jews, Muslims and non-Muslim immigrants from predominantly Muslim countries) were also interviewed. Results from these surveys are presented in chapter 7 of this volume.4

3. Measuring antisemitism

Determining the prevalence of negative attitudes towards Jews in the Norwegian population through the help of a survey is no easy task. The distribution of responses to a specific question will depend not only on the subject matter, but also on the wording and response options provided.5 The research group at CHM therefore decided to use multiple questions to construct indices that combine questions with related content. This way, more reliable measures may be obtained by reducing the impact of random errors, as well as more valid measures of complex phenomena that cannot be captured by a single question.

The indices cover three aspects of antisemitism: an affective dimension of dislike of Jews, a dimension of social distance from Jews, and a cognitive dimension of prejudice where negative characteristics are ascribed to Jews.6 Finally, the three indices are combined in an overall index of antisemitism. These measures were the same in the 2011 and 2017 population surveys. Although there may be some uncertainty regarding the estimated level of antisemitism in each year, since this will depend on the measuring instruments, there will be less uncertainty regarding the direction of change in that level between the two points in time.

3.1 Index for dislike of Jews

The index is mainly based on a question asking how respondents react to the statement “I have a certain dislike of Jews” (Table 4.1). In 2011, three out of four respondents felt that the statement did not fit with their own opinion, 43% not at all, and 34% rather badly.7 In 2017, the corresponding figures were 49% and 33%. The share for the two negative responses taken together fell by 3.7 points, from 11.2% to 7.5%, a significant reduction (1% level). These answers were scored 1 and 2 on the index, and all other answers scored 0.

Table 4.1.

Negative and positive feelings towards Jews (Percent. Population samples)

How well do these statements fit with your own opinion?YearNot at allRather badlyImpos. to answer / NARather wellCompletelyTotal
I have a certain dislike of Jews2011 43.433.911.59.51.7100.0
201748.632.711.35.91.6100.0
I have a particular sympathy for Jews201121.331.720.420.85.9100.1
201722.027.623.320.86.3100.0

The responses to the statement concerning sympathy were more evenly distributed, with a majority that did not find it fitting, down from 53% in 2011 to 50% in 2017, and a quarter of the respondents who did. This question was used to adjust the index score, by assigning the score of 0 on the index for respondents expressing both dislike and sympathy. This contradictory pattern may be a case of response error, but it may also reflect a genuine ambivalence. Feelings can be positive due to, for instance, the particular history of the Jews, yet simultaneously negative due to, for instance, Israel’s policies towards the Palestinians today. Regardless, there may be grounds for disregarding such an ambivalent response pattern when defining dislike of Jews, and only including respondents who only express dislike.

This adjustment reduces the percentage scoring high on dislike of Jews (score 1 or 2 on the 0–2 index) from 11.2% to 9.8% in 2011 and from 7.5% to 6.7% in 2017 (Figure 4.1), compared to the share of respondents expressing dislike (Table 4.1).8

Figure 4.1.

Index for dislike of Jews (Percent. Population samples).

The reduction of 3.1 percentage points in the share of respondents scoring high on the index for dislike of Jews between 2011 and 2017 is significant (1% level).

3.2 Index for social distance from Jews

This index uses questions similar to items from Bogardus’s social distance scale.9 When asked how much they would like to have Jews as neighbours or in their circle of friends, most respondents replied “wouldn’t mind it”. Between 2011 and 2017, the combined share for “dislike a little” or “dislike a lot” fell from 10.6% to 7.0% regarding Jews as neighbours, and from 9.8% to 7.0% for Jews as friends. Both reductions are significant (1% level).

Table 4.2.

Social distance from Jews (Percent. Population samples)

To what extent would you like or dislike:YearLike itWouldn’t mind itNo answerDon’t knowDislike it a littleDislike it a lotSum
Having Jews as neighbours?201113.673.00.02.97.43.2100.0
201713.675.40.23.85.31.7100.0
Having Jews in your circle of friends?201113.972.80.03.46.43.4100.0
201717.970.50.34.35.02.0100.0

For each question the answer “dislike a little” was scored 1 and “dislike a lot” 2 on the index. Top scores are quite rare (Figure 4.2). When the index is dichotomised so as to consider scores 2–4 as a high social distance, the share is 8.5% in 2011, falling to 5.9% in 2017, a reduction of 2.6 percentage points (significant at 1% level).

Figure 4.2.

Index for social distance from Jews (Percent. Population samples)

3.3 Index for prejudices against Jews

The questionnaire included a series of statements about Jews that express stereotypical, generalised, negative images that are commonly found in antisemitic ideas regarding issues such as power, finance and blame. In antisemitism research there are several prejudice indices, which served as a basis for the construction of our index.10 Table 4.3 shows the share of the respondents in 2011 and 2017 who find that a statement fits “completely”, “rather well”, “rather badly” or “not at all” with their own opinion. “Impossible to answer” was also a response option, and the few respondents who did not tick any response alternative are listed under NA (“No answer”) in the table.

The statements are ordered according to how many respondents chose one of the two answers expressing a negative opinion in 2011, varying between 13% and 26%. In 2017 the corresponding proportions range from 8% to 18%. For all statements the percentage expressing agreement is smaller in 2017, with a decrease of between 3 and 8 percentage points. All the changes are statistically significant (1% level).

Table 4.3.

Prejudices against Jews (Percent. Populations samples)

Below is a list of statements that have previously been made about Jews. How well do they fit with your own opinion?YearStatement fits:Sum
Not at allRather badlyImpos. to answer/ NARather wellCompletely
Jews consider themselves to be better than others201115.821.636.319.96.4100.0
201720.023.738.413.64.3100.0
Jews have too much influence on the global economy201117.124.537.516.44.4 99.9
201719.826.041.4 9.83.1100.1
World Jewry is working behind the scenes to promote Jewish interests201117.420.742.915.23.9100.1
201720.921.244.8 9.63.5100.0
Jews have always caused problems in the countries in which they live201127.230.327.911.23.4100.0
201731.031.329.4 6.02.3100.0
Jews have enriched themselves at the expense of others201122.228.135.111.82.8100.0
201723.128.636.7 9.02.6100.0
Jews largely have themselves to blame for being persecuted201138.627.121.710.32.3100.0
201739.430.522.0 6.41.7100.0

For the index of prejudice against Jews, a score of 1 is assigned to the response “fits rather well” and 2 to “fits completely”, giving an additive index ranging from 0 to 12 points (Figure 4.3). The distributions show high proportions for the lowest score (0), telling us that most of the respondents did not find that any of the six negative statements matched their own opinion. This holds for 55% in 2011 and 69% in 2017, an increase of as much as 14 percentage points.

The percentage of respondents scoring in the 10–12 interval was just 1.3 in 2011 and 1.4 in 2017. Above the midpoint of the scale (7–12 points) the percentages were 4.6 in 2011 and 3.3 in 2017. The decrease of 1.3 percentage points is small but significant (5% level).

Figure 4.3.

Index of prejudice against Jews (Percent. Population samples)

On the dichotomised prejudice index, the cut-off point between high and low was set between scores of 3 and 4. This means that as a minimum, two of the six negative statements have been considered to fit with their own opinion. According to this dichotomy, 12.1% of the respondents showed high levels of prejudice against Jews in 2011, falling to 8.3% in 2017, a decrease of 3.8 percentage points (significant 1% level).

3.4 Index of antisemitism

The summary index of antisemitism is an additive index of the three dichotomised sub-indices scored 0 and 1 (Figure 4.4). The vast majority have no high scores on the sub-indices, increasing from 80% to 87% between 2011 and 2017. In some of the analyses that follow, the combined index is dichotomised with a high score on at least two of the three sub-indices defined as a high level of antisemitism. This applies to 7.8% of the population in 2011 and 5.5% in 2017. The reduction of 2.3 percentage points is significant (1% level).

Figure 4.4.

Combined index of antisemitism (Percent)

3.5 Testing the validity of the antisemitism index

We can test whether the index in fact captures what we mean by antisemitism by examining the association between index scores and various opinions where attitudes towards Jews can be expected to create clear differences in the distribution of responses. Figure 4.5 shows such associations, with sharply increasing or decreasing proportions as we move from score 0 to 3 on the combined antisemitism index.

The difference between the two extreme groups scoring respectively 0 and 3 on the index is 65 percentage points regarding acceptance of a Jew as prime minister, 46 percentage points for seeing harassment and violence against Jews as an attack on our society, or as justifiable considering how Israel treats the Palestinians, and 75 percentage points for thinking that Jews exploit Holocaust victimhood for their own purposes. The higher the distance between two extreme groups, the stronger is the relationship between antisemitism and the attitude in question. This pattern gives reason to conclude that the index is a valid measure of antisemitism, actually measuring what was intended.

Figure 4.5.

Validity test for antisemitism index (Percent. Population 2017)*

*N for the first question is reduced from 1,575 to 804 due to a split ballot procedure for this question and the one concerning a Muslim as prime minister (Figure 4.10).

In 2017, the gaps in the distribution of responses for the test questions are larger between scores 0 and 1 than between scores 1 and 2. This means that the group with score 1 more closely resembles the group with score 2 than the group with score 0. This is an argument for using the dichotomy 0 versus 1–3, which would give the percentages of 20.3 in 2011 and 12.4 in 2017 for a high level of antisemitism. Such a dichotomisation would imply a slightly sharper reduction in antisemitism in Norway between 2011 and 2017, with a 6.9 percentage point drop instead of the 2.3 points shown in figure 4.4.11

When validity was tested in the report for the 2011 survey using other test questions, the largest gap in the distributions occurred between scores 1 and 2 on the combined index.12 This was one reason why the dichotomisation of 0–1 versus 2–3 was chosen. It was also seen as reasonable not to use the antisemitism label for respondents who had a high score on just one of the three sub-indices. Both in 2011 and 2017, it is a high score on the prejudice index alone that is most common for those scoring 1 on the combined index (this holds respectively for 41% and 47% in 2011 and 2017). The prejudice index may be more open for discussions regarding the choice of indicators than the other two sub-indices.

In order to get results comparable with those reported for 2011, the following analyses will stick to the dichotomy used in the prior report, with 0–1 versus 2–3 for low versus high level of antisemitism. The same dichotomisation is also used for the summary index of Islamophobia.

Figure 4.6.

Alternative indices of antisemitism (Percent. Population 2017)

There is no denying that the decision of where to draw the line when the sub-indices are dichotomised is also somewhat arbitrary, and will affect the estimated prevalence of antisemitism in Norway.13 The extent of this is seen when we compare the two extremes of a narrow and a broad definition of dislike, social distance and prejudice (Figure 4.6). In the first case, only a top score on the sub-index is considered a high value; in the second case, all index scores above 0.14 The resulting proportions for a score of 0 on the combined index vary between 67% and 97%. A high score (2–3) varies less, from 1% with a narrow to 9% with a broad definition.

The dichotomisation used for the three sub-indices lies between the broad and narrow definitions, with around two-thirds of the index scale defined as a high value. The result for the dichotomised antisemitism index of 5.5% high is slightly closer to the result for the broad than the narrow definition.15

4. Measuring Islamophobia

In the 2017 population survey, negative attitudes towards Muslims were measured using the same kind of indices as for Jews. The questions in the dislike and social distance indices are identical with those in the corresponding indices for Jews. The statements used to measure prejudice are necessarily different, although some of them have content resembling statements in the index for prejudice against Jews.

4.1 Index for dislike of Muslims

A majority of 56% find the statement of dislike as “not at all” or “rather badly” fitting with their own opinion, compared to 30% who see it as “rather well” or “completely fitting” (Table 4.4). The result though, is far more negative for Muslims than for Jews in 2017, where 81% found the dislike statement “not fitting” and only 8% “fitting” (Table 4.1).

Table 4.4.

Dislike of Muslims (Percent. Population 2017)

How well do these statements fit with your own opinion?Not at allRather badlyImpos. to answer / NARather wellCompletelyTotal
I have a certain dislike of Muslims23.132.913.522.57.9 99.9
I have a particular sympathy for Muslims32.332.920.411.82.6100.0

After adjusting for respondents indicating sympathy as well as dislike, the proportion with a high score on the dichotomised index is 28% for Muslims (Figure 4.7), as compared to 7% for Jews (Figure 1).

Figure 4.7.

Index for dislike of Muslims (Percent. Population 2017)

4.2 Index for social distance from Muslims

The questions regarding social distance from Muslims were asked also in 2011. A negative feeling towards Muslims as neighbours (dislike it a little or a lot) was expressed by 28% in 2011 and 26% in 2017. The corresponding results for dislike of having Muslims in circle of friends are 25% and 21% (Table 4.5). These figures are substantially higher than those for Jews, which lie between 7 and 11% (Table 4.2). The reductions in the share of negative answers, which are most pronounced for “dislike a lot”, are small (2.2 and 3.8 percentage points), but the latter is significant (5% level). The trend may appear surprising for some in view of a general impression of a growing scepticism towards immigrants and Muslims in particular among Norwegians. Such a negative trend has not, however, been confirmed by opinion research.16 There also is a positive correlation between the share of immigrants in a local community and positive attitudes towards them, suggesting that part of the explanation for the observed trend towards more positive attitudes is immigration itself.

Table 4.5.

Social distance from Muslims (Percent. Population samples)

To what extent would you like or dislike:YearLike itWouldn’t mind itNo answerDon’t knowDislike it a littleDislike it a lotSum
Having Muslims as neighbors?20116.962.60.02.715.312.6100.1
20178.060.41.44.614.910.8 99.9
Having Muslims in your circle of friends?20119.562.70.03.412.811.7100.1
201713.259.11.75.311 .4 9.3100.0

The social distance index for Muslims has shares of high scores around one-fifth of the population (Figure 4.8), compared to between 8% and 6% for distance from Jews (Figure 4.2). There is a modest decrease of 2.4 percentage points between 2011 and 2017 (significant 5% level).

Figure 4.8.

Index of social distance towards Muslims (Percent. Population samples)

4.3 Index for prejudice against Muslims

The statements used to measure whether the respondents hold negative, stereotypical opinions of Muslims necessarily differ from those in the index for prejudice against Jews.17 The results in Table 4.6 are thus not directly comparable to those of Table 4.3. The share of negative answers for Muslims varies between 29% and 47%, as compared to between 8% and 18% for prejudices against Jews in 2017.

Table 4.6.

Prejudices against Muslims (Percent. Population sample 2017)

Below is a list of statements that have been made about Muslims. How well do they fit with your own opinion?Statement fits:Sum
Not at allRather badlyImpos. to answer/NARather wellCompletely
Muslims largely have themselves to blame for the increase in anti-Muslim harassment10.822.619.130.916.5100.0
Muslims consider themselves morally superior to others9.716.628.627.617.5100.0
Muslims pose a threat to Norwegian culture15.830.014.824.614.8100.0
Muslims do not fit into modern Western society14.431 .817.423.213.2100.0
Muslims want to take over Europe20.123.026.616.613.7100.0
Muslims are more violent than others18.027.525.519.1 9.9100.0

On the index for prejudice against Muslims, nearly 20% of the respondents score above the midpoint of 6 on the scale, and 34% score high (4–12) on the dichotomised index (Figure 4.9). This is far above the corresponding results for prejudice against Jews of 3% and 8% in 2017 (Figure 4.3).

Figure 4.9.

Index of prejudice against Muslims (Percent. Populations sample 2017)

4.4 Index of Islamophobia

For the summary index of Islamophobia made from the dichotomised sub-indices, nearly 60% of the respondents have a low score on all three (Figure 4.10). The rest is evenly split between 1, 2 and 3 high scores, with shares of 13–14%. For the dichotomised index of Islamophobia, 27% have a high score. Once again, the result for Muslims is markedly more negative than for attitudes towards Jews, with a high score of 5.5% in 2017 (Figure 4.4).

Figure 4.10.

Index of Islamophobia (Percent. Population sample 2017)

4.5 Testing the validity of the Islamophobia index

As for the antisemitism index, we have tested the validity of the Islamophobia index by looking at the association between index scores and other attitudes one would expect to be highly correlated with negative attitudes towards Muslims. The first two questions in the test, and to some degree also the third one, are similar in content as those used in the test for antisemitism. As expected, figure 4.11 shows markedly increasing or decreasing proportions as we move from score 0 to 3 on the combined Islamophobia index.

The difference between the two extreme groups is 45 percentage points regarding acceptance of a Muslim as prime minister, 27 percentage points for seeing harassment and violence against Muslim as an attack on our society, 26 percentage points for seeing harassment and violence against Muslims as justifiable considering recent terrorist attacks, and 52 percentage points for thinking that harassment and violence against Muslims would not be a problem if there were fewer Muslim asylum seekers. These differences, although somewhat smaller than the corresponding results for the antisemitism index (Figure 4.5), suggest that the index is a valid measure of Islamophobia.

Figure 4.11.

Validity test for Islamophobia index (Percent. Population 2017)*

* N for the first question is reduced from 1,575 to 771 due to a split ballot procedure for this question and the one concerning a Jew as prime minister.

Even more than for the antisemitism index, how the sub-indices are dichotomised affects the estimated prevalence of Islamophobia in Norway. This becomes clear when we compare the two extremes of a narrow and a broad definition of dislike, social distance and prejudice, as explained earlier in section 3.5. The resulting proportions for a score of 0 on the combined index vary between 35% and 81% (Figure 4.12). A high score (2–3) varies from 8% with a narrow to 34% with a broad definition, as compared to the result of 27% for the index used in our analyses.18

Figure 4.12.

Alternative islamophobia indices (Percent. Population sample 2017).

5. Perception of the prevalence of antisemitism and Islamphobia

How widespread are negative attitudes towards the two minorities perceived to be by members of the Norwegian population? Far more respondents believe negative attitudes towards Muslims to be widespread than negative attitudes towards Jews (Table 4.7). The proportions in 2011 and 2017 respectively answering “fairly” or “very” widespread are 86% and 81% in relation to Muslims and 20% and 19% in relation to Jews.

The change with regard to the perception of negative attitudes toward Jews is too small to be significant. However, at the same time there was an increase in the proportion who believed negative attitudes towards Jews not to be widespread at all, from 7% to 10%, which is significant (1% level). This tendency for the change in the general impression of the prevalence of antisemitism coincides with the actual opinion trend as measured by our indices, all of which, as we have seen, show a modest decrease for negative attitudes towards Jews in the Norwegian population between 2011 and 2017.

For the Muslims, we only have trend data for attitudes regarding social distance, which show a modest decrease (Figure 4.8). As for antisemitism, this is in line with how the opinion climate regarding Muslims is perceived as somewhat less negative in 2017 than in 2011.

Table 4.7.

Impression of the prevalence of negative attitudes towards Jews and Muslims (Percent. Population samples)

How widespread do you think negative attitudes are in Norway today?YearVery wide- spreadFairly wide- spreadImpossible to answerNot very wide-spreadNot wide- spread at allSum
Towards Jews2011 1.718.712.760.1 6.7 99.9
2017 2.416.911.858.810.1100.0
Towards Muslims201120.765.7 3.210.1 0.3100.0
201716.564.3 4.714.0 0.5100.0

Does the public acknowledge the need to combat harassment against these minorities? Most of those who expressed an opinion believe that measures to combat anti-Jewish harassment are needed, increasing from 38% in 2011 to 41% in 2017 (Table 4.8). The corresponding figures with regard to anti-Muslim harassment are higher, but slightly decreasing, from 59% to 56%. The results mean that while twice as many respondents considered it important to combat anti-Jewish harassment as believed negative attitudes towards Jews to be widespread, the pattern is the opposite for negative attitudes towards Muslims. In this case, a larger proportion believed such negative attitudes to be widespread than saw a need to combat anti-Muslim harassment.

Table 4.8.

Need for combating anti-Jewish and anti-Muslim harassment (Percent. Population samples)

Do you see a need to do something to combat harassment in Norway?YearYesNo opinionNo answerNoSum
Against Jews201137.532.50.129.9100.0
201740.731.20.028.1100.0
Against Muslims201159.320.90.019.7100.0
201756.126.10.117.7100.0

There is a strong correlation between respondents’ own attitudes and their impression of the prevalence of negative attitudes in others. High scores on the antisemitism or Islamophobia indices tend to go together with the belief that such attitudes are widespread. In the 2017 survey, 17% of respondents who scored 0 on the combined index for antisemitism believed negative attitudes towards Jews to be very or fairly widespread (most answered “fairly”). In the small group with the top score of 3 on the index, 51% believed negative attitudes towards Jews to be widespread (one-fifth answered “very”). In other words, respondents who themselves are prejudiced towards a certain group tend to think that others are too (Figure 4.13).

Figure 4.13.

Antisemitism and opinion on prevalence of negative attitudes towards Jews in Norway and the need to combat them (Percent. Population sample 2017)

There is also a high correlation between respondents’ own attitudes and their assessment of the need for measures to combat anti-Jewish harassment in Norway: the more negative the attitudes of respondents according to the antisemitism index, the less need they see for such efforts. The pattern suggests that people who are themselves negative would rather promote than combat such attitudes.

Similarly, the results for the Islamophobia index show that the more negative the attitudes of the respondents themselves, the more often they believe such attitudes to be widespread in the general population. In the 2017 survey, most of the respondents in the population sample have the impression that negative attitudes towards Muslims are very or fairly widespread, increasing from 77% of those who scored lowest on the combined index of Islamophobia to 93% of those who scored highest (Figure 4.14). The proportion that answered “very widespread” rose from 12% to 39%.

Figure 4.14.

Islamophobia and opinion on prevalence of negative attitudes towards Muslims in Norway and the need to combat them (Percent. Population sample 2017)

A similar pattern as for antisemitism also exists with regard to the relation between respondents’ own attitudes and their opinion on whether it is necessary to combat anti-Muslim harassment in Norway. The more negative the attitudes of respondents according to the index of Islamophobia, the less often they see the need for measures to combat anti-Muslim harassment.

6. The relationship between antisemitism and Islamophobia

Are antisemitism and Islamophobia related phenomena, or attitudinal opposites? Is it a matter of both–and or of either–or when it comes to such attitudes? The former is conceivable if xenophobia is a dominant influence behind these attitudes. The latter might be expected if the Israeli-Palestinian conflict plays a decisive role in the attitude formation, and individuals develop positive attitudes towards the party they sympathise with and negative attitudes towards its opponent.

That there is a tendency for antisemitism and Islamophobia to coincide in individuals is reflected by the correlation between these two indices and the two sets of sub-indices. All coefficients (Pearson’s r) are positive in the population sample in 2017 (Table 4.9). For the two summary indices the coefficient equals 0.24, a clearly significant though not very strong correlation.19

Table 4.9.

Correlations between indices for negative attitudes towards Jews and Muslims (Pearson’s r (p <0.001 for all). Populations sample 2017).

Negative attitudes: Jews Negative attitudes: Muslims
Prejud.DislikeDistan.Antisem.Prejud. DislikeDistan.Islamo.
Prejudice against Jews10.560.470.780.270.190.210.22
Dislike of Jews0.5610.530.800.180.160.170.16
Social distance towards Jews0.470.5310.740.230.200.330.25
Antisemitism0.780.800.7410.250.210.270.24
Prejudice against Muslims0.270.180.230.2510.680.700.85
Dislike of Muslims0.190.160.200.210.6810.660.84
Social distance towards Muslims0.210.170.330.270.700.6610.80
Islamophobia0.220.160.250.240.850.840.801

Table 4.10 presents three versions of the relationship between the two dichotomised summary indices. The left part shows that the likelihood for scoring high on Islamophobia is far greater for people with a high level of antisemitism than for people with low. The difference is 30 percentage points. Correspondingly, the middle part of the table shows that scoring high on Islamophobia increases the likelihood of having antisemitic attitudes. The difference is 8 percentage points.

Table 4.10.

The relationship between antisemitism and Islamophobia (Percent. Population 2017)

IslamophobiaAntisemitism%-d H-LIslamophobiaAntisemitismSumIslamophobiaAntisemitismSum
LowHighLowHighLowHigh
High255530High8911100High24.03.027.0
Low7545–30Low973100Low70.52.573.0
Sum1001000%-d H-L–880Sum94.55.5100.0

The pattern in Table 4.10 means that the combinations high–high and low–low for the antisemitism and Islamophobia indices will occur more frequently than chance would predict. This is shown to the right in the table. Here, the distribution of the respondents on the two dichotomised indices is shown as percentages of the Grand Total. A majority of 70.5% of all respondents score low on both indices, while 3% score high on both. Antisemitism alone is found in 2.5% of the sample, while Islamophobia alone is found in 24%.

If the responses had been distributed in the cells of the table randomly (by drawing lots), and in such a way that we kept the marginal distributions for the two indices (94.5–5.5 and 73–27), the proportion that fell in the high–high or the low–low cells would be 1.5 percentage points lower in each cell (resulting in 1.5% instead of 3% located in the high–high cell, for example). Correspondingly, the proportion in each of the two cells with a low value on one index and a high on the other, would be 1.5 percentage points higher.

This shows that there is a tendency for antisemitism and Islamophobia to occur in combination. They are, in other words, related attitudes rather than opposites.20 It is clear, however, that they also do occur alone, especially in the case of Islamophobia, since negative attitudes towards Muslims are far more widespread in Norway than antisemitism according to our measures.

7. Possible causes of antisemitism and Islamophobia

It is difficult to draw causal conclusions based on non-experimental survey data. What the data can show are statistical correlations, but these do not necessarily reflect causal influence. Correlations may be spurious, brought about by prior causal variables affecting both variables in question. This section will attempt to reveal non-causal association through analyses where such variables are controlled for. However, one can never be absolutely sure that such a control will cover all the relevant variables. Another problem is causal direction, i.e. in which direction an influence between the variables flows. For example, this can be difficult to know in the case of the relationship between attitudes towards Jews and views on the Israeli-Palestinian conflict. In many cases, it is reasonable to assume that an influence will work both ways.

Which factors may influence peoples’ attitudes towards Jews or Muslims? We will look at how attitudes vary between groups defined by gender, age, and education, as well as religiosity, opinions on the Israeli-Palestinian conflict, xenophobia, and scepticism towards immigrants in Norway.

The aim of the analyses is to form a picture of what may have contributed to individuals in the Norwegian population developing negative attitudes towards Jews or Muslims. For this purpose, separate indices were constructed for opinions on the Israeli-Palestinian conflict, xenophobia, and attitudes towards immigrants.

7.1 Opinion on the Israeli-Palestinian conflict

Two to three times as many respondents in the population sample support the Palestinians in the Israeli-Palestinian conflict as support Israel (Table 4.11). Very few choose the extreme alternatives “support solely”, and more than half of the respondents refrain from expressing an opinion. The results from the 2017 survey show a slight decrease since 2011 in the proportion that supports the Palestinians and an increase in the proportion not taking sides. In the subsequent analyses, the two categories at either end of the spectrum were combined, as were the two categories in the middle expressing no support for either side, thereby reducing the number of values for the variable from eight to five.

Table 4.11.

“People have conflicting views on the conflict between Israel and the Palestinians. Which side do you support most?” (Percent. Population samples)

YearSolely IsraelMostly IsraelTo some extent IsraelNeitherImpossible to answer/NATo some extent PalestiniansMostly PalestiniansSolely PalestiniansSum
20111.36.84.730.320.812.721.12.2 99.9
20172.16.74.531.922.510.518.33.6100.1

Table 4.12 shows that statements expressing positive positions on Israel (the first two) received less support than those expressing positive positions on the Palestinians (last two).21 Norwegians have more faith in the sincerity of the Palestinian than the Israeli leaders when it comes to solving the conflict. The distribution of responses to the pro-Israeli statements was quite similar in both surveys, while the pro-Palestinian statements received slightly less support in 2017.

The content of the two remaining statements (the third and fourth in Table 4.12) is critical of Israel.22 Around one-third of the respondents answered fits “rather well” or “completely” to the statement “Israel treats the Palestinians just as badly as the Jews were treated during World War II”, the proportion being slightly smaller in 2017 than in 2011. The statement “As long as the State of Israel exists there can be no peace” was supported by 20% in 2017 and 16% in 2011.

Table 4.12.

Opinions regarding the parties in the Israeli-Palestinian conflict (Percent. Population samples)

How well do these statements on the Middle East conflict fit with your own opinion?YearStatement fits:Sum
Not at allRather badlyImpos. to answer/NARather wellCompletely
Israel's leaders genuinely want to find a solution to the conflict201112.932.134.216.64.2100.0
201710.031.237.017.84.1100.1
Israel is at the forefront of the war on Islamic terrorism201112.420.446.915.94.5100.1
20178.020.951.914.74.5100.0
As long as the State of Israel exists there can be no peace201124.923.735.611.74.1100.0
201713.220.945.615.84.6100.1
Israel treats the Palestinians just as badly as the Jews were treated during WW2201111.521.029.429.19.0100.0
20179.920.537.425.46.9100.1
Both the Israelis and the Palestinians are entitled to a state of their own20112.53.717.427.848.6100.0
20172.24.423.130.939.5100.1
The Palestinian leaders genuinely want to find a solution to the conflict20116.121.834.431.66.2100.1
20177.319.240.428.25.0100.1

A factor analysis23 of the six statements resulted in three dimensions, each with a pair of the statements. When the responses for each statement are coded from 0 to 4, this gives three additive indices with scores ranging from 0 to 8, called pro-Israeli attitudes (statements 1 and 2), anti-Israeli attitudes (statements 3 and 4), and pro-Palestinian attitudes (statements 5 and 6). Table 4.13 shows the distribution on the indices and how they are dichotomised. The scale is divided just above the midpoint so that scores of 5 to 8 are defined as high values on the index.

Table 4.13.

Indices for opinions on the Middle East conflict (Percent. Population samples)

IndexYearIndex scoreSumHigh 5-8
012345678
Pro-Israeli attitudes20115.27.813.115.539.49.36.81.81.2100.119.0
20173.05.815.214.640.810.36.62.31.4100.020.6
Anti-Israeli attitudes20117.58.712.812.833.512.17.83.81.1100.124.8
20175.25.412.211.238.712.69.33.42.0100.027.2
Pro-Palestinian attitudes20110.71.02.45.224.820.919.720.25.3100.266.0
20171.21.33.05.328.819.321.815.73.8100.260.5

The proportion with a high value remained stable at around 20% from 2011 to 2017 for the pro-Israeli index. Around a quarter of the population sample had a high value on the index for anti-Israeli attitudes, with an insignificant increase from 25% in 2011 to 27% in 2017. The proportion of respondents on the lower end of the scale (scores 0–3), not supporting the anti-Israeli statements, decreased from 42% to 34%. Most of the respondents are located above the midpoint on the pro-Palestinian index, with 66% in 2011 and 60.5% in 2017. Both of these changes are significant at 1% level.

In addition to being dichotomised in multivariate analyses, the indices are trichotomised in some tables. Then a low value will denote scores 0–2, a medium value 3–5 and a high value 6–8.

7.2 Xenophobia

Earlier we presented the attitude towards social contact with Jews and Muslims, defined by whether respondents would like or dislike having them as neighbours or friends. An index of social distance was constructed by assigning 1 point for the response “would dislike it a little” and 2 points for “would dislike it a lot” for each of the two types of contact. Table 4.14 shows the distribution on a similar index of social distance towards Roma, Somalis and Poles. The first two groups in particular stand out with respect to a high level of scepticism in the population sample. The proportion with high scores (2–4) in the 2017 survey is 44% for Roma, 27% for Somalis and 8% for Poles. For the purpose of comparison, the score for Americans was 4% and for Catholics 3%, and, as already presented, 21% for Muslims and 6% for Jews.

Table 4.14.

Indices of social distance towards Roma, Somalis and Poles (Percent. Population samples)

Index of social distance:YearIndex scoreSumHigh (2-4)
01234
Towards Roma (Gypsies)201136.819.719.310.513.7100.043.5
201741.713.919.65.419.4100.044.4
Towards Somalis201152.016.216.16.6 9.2100.131.9
201762.211.112.42.811.6100.126.8
Towards Poles201180.48.38.00.8 2.4 99.911.2
201787.15.25.60.6 1.5100.0 7.7

Table 4.14 shows stability between 2011 and 2017 in negative attitudes (score 2–4) regarding social contact with Roma and significantly less scepticism towards contact with Somalis and Poles (1% level).24

Could reluctance to have contact with Jews or Muslims be part of a more general scepticism towards foreigners, or xenophobia, as it is also known? To measure xenophobia, we use an additive index of the total scores for the three groups in Table 4.14. With three indicators scored 0 to 4, the result is an index ranging from 0 to 12. If a high level of xenophobia is defined as a score above the midpoint on the scale (7–12), we find this in 15% and 13% of respondents in 2011 and 2017 respectively.

Table 4.15.

Xenophobia index based on social distance towards Roma, Somalis and Poles (Percent. Population samples)

YearIndex scoreSumHigh(7-1 2)
0123456789101112
201132.012.512.08.89.84.95.32.85.72.52.10.61.0100.014.7
201737.910.613.94.710.33.06.12.65.80.92.60.21.299.813.4

7.3 Scepticism towards immigrants

The respondents were asked about their views on the economic and cultural consequences of immigration. The questions were presented in the form of a discussion between two people, A and B, and the respondents asked to indicate with whom they agreed most (Table 4.16).25

Majorities of 54% (2011) and 57% (2017) believe that the effect of immigration on Norwegian culture is positive, supporting statement A. Around a quarter of the respondents chose the negative statement of a “threat”. The view concerning the economic effects is not quite as positive, but there is a clear trend towards less scepticism. In 2011, the two alternatives were chosen by 37% each, while in 2017 this had changed to 31% for the “exploit” alternative (A) as opposed to 44% for “contribute” (B).

Table 4.16.

Attitudes towards immigrants (Percent. Population samples)

Two people are discussing the possible effects of immigrants from other cultures arriving in Norway. With whom do you agree most, A or B?
A says: Immigrants contribute to greater cultural diversity in Norway, introducing new and exciting food, music, art, etc.A says: Immigrants want to exploit our welfare system and enjoy benefits which they played no part in creating.
B says: Immigrants’ ways of life don’t fit into Norwegian society. Their foreign customs are problematic for those around them and could threaten Norwegian culture.B says: Immigrants are hard-working, diligent people who make a valuable contribution to the Norwegian economy and working life.
Year A Diversity Imposs. to choose /NA B Threat Sum A Exploit Imposs. to choose /NA B Contribute Sum
201154.220.925.0100.137.125.937.1100.1
2017 56.818.424.8100.031.025.443.6100.0

An index of scepticism towards immigrants was created by assigning a score of 0 for a positive response, 1 for not expressing an opinion, and 2 for a negative response. When the scores of 3 or 4 on the index are regarded as high values, 31% of the sample is classified as being sceptical towards immigrants in 2011 and 29% in 2017, a reduction too small to be significant. The proportion expressing two positive attitudes (score of 0) increased from 31% in 2011 to 38% in 2017 (significant 1% level).

Table 4.17.

Index of scepticism towards immigrants (Percent. Population samples)

YearIndex scoreSumHigh (3-4)
01234
201130.717.421.411.519.1100.130.6
201738.214.918.210.917.9100.128.8

7.4 Possible explanations for negative attitudes towards Jews and Muslims

When the respondents are grouped according to social characteristics or opinions measured by the indices discussed above, variations in the incidence of antisemitism or Islamophobia between the groups may provide clues as to what stimulates development of such attitudes. The dichotomised summary indices, where a high value denotes a high score on at least two of the three sub-indices, are used. Table 4.18 shows how the proportion of respondents displaying high levels of antisemitism or Islamophobia according to this definition varies between different groups in the population.

The incidence of both antisemitism and Islamophobia is higher among men, among older people, and among people with lower levels of education. Belief in God and regarding religion as important in one’s life show no clear correlations with antisemitism or Islamophobia in the general population, though the proportion displaying high levels of Islamophobia among those who answered “yes” to the question about belief in God is larger than for those who answered “no” (significant 1% level).

Table 4.18.

Variation in antisemitism and Islamophobia (Percentage with high scores on the combined indices. Population sample 2017)

VariableValues (index scores)High antisemitismHigh IslamophobiaPercent of sampleN (=100%)
GenderFemale32050786
Male83450789
Age18–29 years21118282
30–44 years52728434
45–59 years73027417
60+ years63428442
EducationUniversity level31832506
Lower731681069
Belief in GodYes63134528
Not sure42825398
No62341649
Importance of religionVery important43112104
Fairly important 52928235
Neither63236304
Not very important43818149
Not important at all732755
Support for parties in Middle East conflictSolely/mostly Pal. 121322345
To some extent Pal.61811165
Neither /No opinion32854856
To some extent Israel547570
Solely/mostly Israel2609138
Pro-Israeli attitudesStrong (6–8)35210162
Medium (3–5)424631035
Weak (0–2)112623378
Anti-Israeli attitudesStrong (6–8)223615231
Medium (3–5)32366984
Weak (0–2)23424360
Pro-Palestinian attitudesStrong (6–8)72041650
Medium (3–5)43153840
Weak (0–2)945585
XenophobiaNone (0)2538597
Weak (1–2)21525387
Some degree (3–4)73615236
Medium (5–6)8559144
Strong (7–12)198313211
Scepticism towards immigrantsNone (0)3338602
Weak (1 )41215234
Medium (2)42918286
Quite strong (3)44211171
Strong (4)138018282
All5271001575

Which side the respondents support in the Middle East conflict and what opinions they hold on the conflict clearly correlate with antisemitism and Islamophobia in the expected direction. The exceptions are that the correlation between pro-Palestinian attitudes and antisemitism is weak, and that strong anti-Israeli attitudes go together with Islamophobia. The first finding indicates that having pro-Palestinian attitudes is not necessarily a result of antisemitism.26 The second finding may be a result of xenophobia and scepticism towards immigrants stimulating both antisemitism and Islamophobia. These attitudes have clear correlations with antisemitism and, in particular, Islamophobia.

In terms of incidence of high levels of Islamophobia, the difference between the groups at the extremes of the indices of xenophobia and scepticism towards immigrants is almost 80 percentage points. It could be questioned whether there is any merit in considering scepticism towards foreigners or immigrants on the one hand and antisemitism or Islamophobia on the other as separate phenomena that may influence each other, as we have done here, or whether they should instead be considered as different aspects of the same phenomenon, a syndrome that has been called group-focused enmity.27

Several of the variables in Table 4.18 are correlated. For example, individuals displaying high levels of xenophobia will often also be sceptical towards immigrants. In order to see what a characteristic in itself means for negative attitudes towards the minorities, the groups to be compared must be made equal with respect to the other variables through a multivariate analysis were these variables are included. This can be done by means of a multivariate regression analysis (Table 4.19). Since importance of religion according to Table 4.18 did not correlate with antisemitism or Islamophobia, this variable is omitted from the analysis.

To ease comparison of the importance of the different explanatory variables, they are dichotomised in the multivariate analysis. The exception is which side respondents supported in the Israeli-Palestinian conflict, which is represented by two dummy variables, with not taking sides as reference group. With the dichotomised indices of antisemitism and Islamophobia, coded 0 for low value and 1 for high value, as dependent variables, the linear regression coefficients equal proportion differences. When multiplied by 100 as shown in the table, the coefficients can be interpreted as percentage differences.28 The bivariate association between gender and antisemitism in Table 4.19 (−4.9) for example corresponds to the difference in percentage points between women and men in Table 4.18 (3–8 = –5).

The variables in the table are divided into two categories. The first contains the social background variables and belief in God, the second opinion on the Israeli-Palestinian conflict, xenophobia and scepticism towards immigrants. The latter group of variables lie closer to antisemitism and Islamophobia in the causal chain, and can be considered as intervening variables producing an indirect effect between the first group of variables and negative attitudes towards Jews and Muslims. They represent potential mechanisms that may explain the correlation between them.

It could also be possible that the influence works in the opposite direction; for instance, that Islamophobia leads to scepticism towards immigrants, or that it works both ways, meaning that these phenomena stimulate each other. This is impossible to determine with the available data, making the causal interpretation of the effects uncertain.

The multivariate analysis is performed in two stages. In the first, the social background variables and belief in God are included. Changes in the bivariate correlation for a variable show how much of this correlation can be explained by the other variables in the group. For some, such as gender and age, this part of the association will be indirect effects. For others, it may also be a case of spurious (non-causal) association caused by variables in the group prior to them in time. In the second stage, all the variables are included, and the remaining association constitutes the direct effect of the variable in question, given the variables included in the model and its assumptions of causal direction.

Table 4.19 shows that when we remove differences between women and men with regard to the other variables, the gender difference for antisemitism is reduced, but only marginally (from –4.9 to –4.4 percentage points). For Islamophobia, however, the effect of gender is radically reduced when controlled for all other variables (from –14.2 to –5.1). This can largely be ascribed to the clear gender differences with regard to xenophobia and scepticism towards immigrants evident in the correlations in Table 4.20. The fact that these correlations are negative indicates that women – who are assigned high value on the gender variable – have lower incidences of such attitudes than men. Table 4.19 shows that both xenophobia and scepticism towards immigrants have a strong effect on the likelihood of scoring high on Islamophobia, and thereby transmit a negative indirect effect between gender and Islamophobia according to the model.

The results of the final multivariate analysis show that xenophobia has the strongest effect on the antisemitism index, followed by anti-Israeli attitudes and supporting the Palestinians in the Middle East conflict, while being a woman reduces the chances for a high score.

Xenophobia has strongest effect on the level of Islamophobia, closely followed by scepticism towards immigrants, which had little effect on the incidence of antisemitism. Supporting Israel in the Middle East conflict increases the chance for a high level of Islamophobia. Supporting the Palestinians reduce the chances, and so does being a woman or young in age.

Table 4.19.

Bivariate and multivariate regression analysis with dichotomised indices of antisemitism and Islamophobia as dependent variables (Regression coefficients multiplied with 100. Population sample 2017)*

VariableHigh value (index scores)AntisemitismIslamophobia
BivariateMultivariateBivariateMultivariate
GenderFemale–4.9–4.8–4.4–14.2–14.0–5.1
Age–44 years–2.4(–2.2)(–0.5)–11.5–9.9–5.1
EducationUniversity–3.5–2.9(–1.7)–12.7–10.4(–0.8)
Belief in GodYes(0.5)(0.3)(1.9)6.14.9–2.6
Israeli-Pal. conflict (2 dummy var.) (Refer. group: Do not take sides)Support Israel(–2.6)(–1.3)33.013.9
Support Palest.6.65.4–18.7–8.0
Pro-Israeli attitudesStrong (5–8)(–2.6)(–1.9)12.7(0.6)
Anti-Israeli attitudesStrong (5–8)12.610.35.05.5
Pro-Palestinian attitudesStrong (5–8)(0.3)(–1.6)–11.1(–0.3)
XenophobiaStrong (7–12)15.914.664.139.9
Scepticism towards immigrantsStrong (3–4)6.2(1.0)54.436.4
Explained variance (adjusted R squared)0.0160.1300.0560.419

* In brackets: Not significant (5% level).

Education has a clear bivariate correlation with Islamophobia, which shows little change when controlled for other social background variables, but disappears when controlled also for attitudinal variables. Thus, according to our analysis, education does not have a direct effect, but rather an indirect one, primarily via xenophobia and scepticism towards immigrants. Such attitudes are less common among people with university or university college education (Table 4.20).

Table 4.20.

Correlation matrix for the independent variables (Pearson’s r. Population samples 2011 and 2017 combined)

VariablesHigh valueGenderAgeEducationXenophobiaImmigrant scepticismSupport IsraelSupport PalestiniansPro-Israel attitudesAnti-Israel attitudesPro-Palest. attitudes
GenderFemale1–0.040.06–0.09–0.15–0.15–0.01–0.060.04–0.03
AgeAge 18–44–0.0410.05–0.06–0.01–0.08–0.07–0.08–0.12–0.14
EducationUniversity0.060.051–0.11–0.17–0.030.11–0.04–0.070.03
XenophobiaHigh–0.09–0.06–0.1110.410.09–0.120.060.05–0.12
Immigrant scepticismHigh–0.15–0.01–0.170.4110.1 6–0.20.070.03–0.16
Israeli-Pal. conflictSupport Israel–0.15–0.08–0.030.090.161–0.280.4–0.18–0.17
Israeli-Pal. conflictSupport Palest.–0.01–0.070.11–0.12–0.2–0.281–0.190.270.31
Pro-Israeli attitudesHigh–0.06–0.08–0.040.060.070.4–0.191–0.070.03
Anti-Israeli attitudesHigh0.04–0.12–0.070.050.03–0.180.27–0.0710.14
Pro-Pal. attitudesHigh–0.03–0.140.03–0.12–0.16–0.170.310.030.141

An intuitive and perhaps more easily understandable way of documenting the effects of these variables on antisemitism or Islamophobia is through a tabular analysis. However, there is a limit to how many characteristics that can be examined simultaneously in order to avoid getting too few respondents in the cells of the table. Tables 4.21 and 4.22 use two of the independent variables shown by the multivariate analysis to have the greatest effect on antisemitism and Islamophobia, namely xenophobia and opinion on the parties in the Israeli-Palestinian conflict. Limiting the number of independent variables to two allows the use of five values for each in the table, instead of the crude dichotomy used in the regression analysis.

With two independent variables of five categories each, we get 25 combinations, which provide a wide variation in the proportion with a high level of antisemitism (Table 4.21). The percentage ranges from 0 in the bottom left-hand corner for respondents with no xenophobia who support Israel, to 52 in the upper right-hand corner for respondents with high levels of xenophobia who support the Palestinians. Between these extremes, the percentage with high antisemitism gradually increases in a pattern that follows the main diagonal of the table.

Table 4.21.

Percent high antisemitism depending on xenophobia and opinion on the Israeli-Palestinian conflict (Population samples 2011 and 2017 combined)

XenophobiaWhich party supported in the conflictDifference
Solely/Mostly IsraelTo some extent IsraelBoth/none To some extent PalestiniansSolely/ Mostly Palestinians
Very high 5.911.819.633.252.046.1
High2.03.86.95.020.618.6
Medium0.02.42.17.722.122.1
Low1.50.01.57.47.56.0
Very low0.00.01.10.74.74.7
Difference5.911.618.532.547.341.4

Table 4.22 for Islamophobia is set up in the same way as Table 4.21. Since the correlation with opinion on the parties in the conflict has the opposite sign as for antisemitism, the proportions increase from the bottom right-hand corner to the upper left-hand corner along the bi-diagonal. The variation ranges from 2% among respondents with no xenophobia who strongly support the Palestinians, to 91% among respondents with very high levels of xenophobia who strongly support Israel.

Table 4.22.

Percentage of high Islamophobia depending on xenophobia and opinion on the Israeli-Palestinian conflict (Population sample 2017)

XenophobiaWhich party supported in the conflictDifference
Solely/ Mostly IsraelTo some extent IsraelBoth/noneTo some extent PalestiniansSolely/ Mostly Palestinians
Very high90.992.583.772.666.724.2
High87.456.549.340.257.230.2
Medium67.744.434.533.320.347.4
Low42.114.115.47.16.335.8
Very low17.322.15.81.21.915.4
Difference73.670.477.971.464.88.8

In Table 4.21 for antisemitism, the distance between the extreme groups is slightly greater for xenophobia (columns) than for opinion on the Middle East conflict (rows), with mean differences of 23.2 and 19.5 percentage points respectively. This applies even more so for Islamophobia (Table 4.22), with mean differences of 71.6 and 30.6 percentage points. The pattern testifies to the importance of xenophobia – a general scepticism towards foreigners – for the development of negative attitudes towards Jews and, in particular, Muslims.

8. Explaining trends in antisemitism

The two population surveys have shown a reduction in the share of respondents with high scores on the antisemitism index in Norway in 2017, down from an already low level in 2011. The question raised in this section is what can explain such a trend. It will be addressed first by looking at the role played by generational replacement versus individuals changing their opinion, secondly by looking at changes in the variables that, according to the analysis in the previous section, affect antisemitism.

8.1 Generational replacement or period effects

Table 4.18 showed that a high level of antisemitism is three times more common in the oldest than in the youngest age group: 6% versus 2%. The same holds for Islamophobia, with 34% versus 10% for the 60+ years old compared to the young of 18–29 years old.29 Does this reflect a life-phase effect, where people grow more sceptical toward strangers and foreign cultures as they age? Or is it a sign that new generations have developed attitudes that differ from those of older generations due to changed circumstances during adolescence, the formative years for the values of an individual?30 If the latter is the case, this means that generational replacement over time will change the population opinion climate. The question is to what extent replacement explains the reduced antisemitism in Norway, or to what extent this trend is a result of individuals present through the whole period changing their opinion, so-called period effects.

These are questions addressed by cohort analysis, where cohorts (generations) are followed over time to see whether they have stable characteristics that differ between them, giving rise to generation replacement effects.31 Or does the opinion of the individuals within the cohorts change over time due to the impact of historical events or processes, producing so-called period effects that change popular opinion? This is in contrast to individual changes related to life phase, which will not affect overall opinion unless the age distribution of a society changes markedly.

Table 4.23 is a standard cohort matrix, with age groups six years wide placed along the left margin, and the two points of observation, six years apart, placed over the columns. In this way, we may follow a cohort by reading the table diagonally as indicated by the shading. The tendency within the cohorts, as captured by the mean of their changes, is a reduction of the percentage of high antisemitism within the cohorts of 1.7 points. It is unlikely that this should be a life-phase effect, since the tendency is away from, rather than towards, the more negative attitudes of older people. The reduction is a little less than the change for the population as a whole between 2011 and 2017 of –2.3 percentage points, indicating that generational replacement has also played a role.

Table 4.23.

Percentage with high antisemitism (Standard cohort matrix, population samples)

AgeYearBirth cohortCohort changeN=100% (weighted)Simplified matrix
201120172011201720112017
18–238.03.01988–9317794Stayers Time 1 7.6G.in 3.0
24–292.11.81988–93–6.2145188Stayers Time 2 5.6
30–355.35.91982–873.8148199
36–416.15.21976–81–0.1156155
42–477.74.31970–75–1.8168153
48–535.17.71964–690.0157142
54–5916.77.91958–632.8143203
60–659.05.71952–57–11 .0239154
66–718.84.81946–51–4.2143168Gen. out 9.4
72–7711.99.41940–451.635102
78–(8.9)(0.0)1934–39917
All7.85.5Change –2.3Mean –1.715221575

The cohort patterns in Table 4.23 are varied, which to some extent may be a result of random errors due to small bases for the percentages in the cells. An alternative to the full matrix is a simplified version where we distinguish between cohorts taking part in the replacement process – the out-going and the in-coming generation – and cohorts present at both times (called stayers at time 1 and time 2). The in-generation is respondents 18–23 years in 2017, who were too young to be part of the sample from the adult population in 2011. The members of the out-generation are not as easily defined. It should be those members of the adult population in 2011 that have died between 2011 and 2017. They would have come from several age groups, but predominantly the oldest ones. In the analysis, we let the age groups 66 years and older represent the out-generation.

The difference between the in- and the out-generation in the prevalence of high antisemitism is 3.0 – 9.4 = –6.4, and the change between 2011 and 2017 for the “stayers” is 5.6 – 7.6 = –2.0. Table 4.24 also gives the results for the three sub-indices. The differences between the in-coming and the out-going members of the population are larger than the changes in opinion for those present at both points in time.32 This especially holds for prejudice.

Table 4.24.

The importance of generation and period effects for negative attitudes towards Jews (Percent. Population samples)

Change over time in popular opinion due to:Relevant differences:Indices for negative attitudes towards JewsN (=100%)
DislikeDistancePrejudiceAntisem.
Change of members of population (generational replacement)In: 18–23 years 20172.03.93.93.094
Out: 66+ years 20119.110.819.69.4187
Difference In – Out–7.1–6.9–15.7–6.4
Change of opinion among stayers between 2011 and 20172017: 24+ years7.06.08.55.61481
2011: 18–65 years9.98.211.07.61335
Change Time2–Time1–2.9–2.2–2.5–2.0
Resulting population change from 2011 to 2017–3.1–2.6–3.8–2.3

When the actual changes from 2011 to 2017 for the total samples lie close to the period effects, the reason is that the group of stayers is so much larger than the groups being exchanged. This is due to the short time span of six years. Over a longer period, the generational replacement would involve larger shares of the population and contribute more to the population trend, but in the present six-year period not more than 6% are newcomers in 2017.

8.2 Changes in variables affecting antisemitism and Islamophobia

Why do the in- and out-going generations between 2011 and 2017 differ in attitudes towards Jews? What has caused a net shift in the attitudes of individuals in the cohorts present at both points in time? This may have to do with changes in the independent variables that, according to the analyses in section 7.4, have an effect on antisemitism. For this to be the case, the variables – in addition to affecting antisemitism – must have changed in the “right” direction in this time period (i.e. show a decline for the value that increases the likelihood of antisemitism or an increase for the value that reduces this likelihood).

The criterion of change in incidence excludes variables such as gender and age, where the composition of the population will not have changed much during the time period in question. It also excludes variables with negligible direct effect on antisemitism in the multivariate analysis in Table 4.19, such as religiosity. This leaves us with the variables in Table 4.25.

A variable’s contribution to changes in the incidence of high levels of antisemitism equals how much it has changed multiplied by its effect on antisemitism. It turns out that the changes in incidence in particular are so negligible that this contribution amounts to only a few tenths of one per cent. The greatest contribution – for anti-Israeli attitudes – even has the “wrong” sign; the trend towards slightly higher incidence of such attitudes should have contributed to more, not less, antisemitism. The result, when contributions for all the variables are added up, is –0.2 percentage points.

Considering the actual decline of 2.3 percentage points, we must conclude that the changes in the variables in Table 4.25 cannot explain the decline in antisemitism in Norway between 2011 and 2017. In order to understand the background for this development, we must look for trends or events during this period that are not captured by these variables. One possibility might be increased media and political attention to antisemitism as a social issue during this period, generated by terrorist attacks against Jews in Europe, among other things.

Table 4.25.

Effect of changes in independent variables on the trend in antisemitism (Percent. Population samples)

VariableHigh values (index scores)IncidenceDirect effect on antisem. (2017)Change x Direct effect
20112017Change
EducationUniversity28.432.13.7–1.7–0.06
Middle East conflictSupport Palest.36.032.4–3.65.4–0.19
Anti-Israeli attitudesStrong (5–8)24.827.22.410.30.25
XenophobiaStrong (7–12)14.713.4–1.314.6–0.19
Scepticism towards immigrantsStrong (3–4)30.628.8–1.81.0–0.02
AntisemitismHigh (2[g] 3)7.85.5–2.3Total:–0.21

For Islamophobia we only have data for social distance in 2011, which show a small reduction in 2017. For the other indices, the actual amount of change is not known. Although modest in size, the reduction in xenophobia and scepticism towards immigrants shown in Table 4.25 may have contributed to a reduced incidence of all kinds of negative attitudes due to the strong effects these variables have on Islamophobia (Table 4.19).

9. Conclusion

The level of negative attitudes towards Jews in Norway is low and declining, according to our measures. In the 2017 survey, 6.7% scored high on the index of dislike, a reduction of 3.1 percentage points from 2011. On the index for social distance, 5.9% scored high in 2017, down 2.6 points from 2011. The percentage scoring high on the index for prejudice was 8.3 in 2017, down 3.8 points from 2011. The summary index of antisemitism showed that 5.5% had a high score on at least two of the three sub-indices in 2017, a reduction of 2.3 points from 2011.

The corresponding levels of negative attitudes towards Muslims in 2017 are much higher. For the dislike index, 27.7% score high, for social distance 19.6%, for prejudice 34.1% and for the summary index of Islamophobia 27.0%. The only index where we have results also for 2011 regarding Muslims – social distance – shows a reduction in high scores of 2.4 percentage points.

In 2017, negative attitudes towards Jews were perceived to be very widespread by just 2.4% of the respondents. If we add fairly widespread, the result is 19.3%, a figure which seems high compared to our results for measures of actual popular opinion. The same holds for the perception of negative attitudes towards Muslims, which is 16.5% for very widespread and 80.8% when we add fairly widespread. The perception of the opinion climate regarding Muslims has become less negative from 2011 to 2017 (5.6 percentage points for the two answers combined). Regarding Jews there is a tendency in the same direction, but this is too small to be significant.

There is a clear tendency that the more negative the attitudes of a person towards Jews or Muslims are, the more likely it is that he or she will perceive the general opinion climate as negative, and the less likely it is that an effort to combat harassment against these minorities is seen as necessary.

It turns out that there is a tendency for negative attitudes towards the two minorities to go together. Accordingly, antisemitism and Islamophobia can be seen as related phenomena rather than opposites, with xenophobia as the most important stimulating factor. In addition negative attitudes towards Israel go together with antisemitism, and scepticism towards immigrants with Islamophobia.

In the years to come, will we see a continuation of the trend towards less negative attitudes towards these minorities? The development for antisemitism and Islamophobia in Norway will depend upon generational replacement as well as individuals changing their opinion influenced by current events. Judging from the present generational differences, the first process may be expected to stimulate a continued gradual reduction in the prevalence of negative attitudes towards Jews as well as Muslims, among other things as a result of an increasing level of high education in the new generations.

The effect of historical events on attitudes is more uncertain. Up until now the growing number of immigrants in Norway seems to have affected the attitude of Norwegians towards Muslims positively, but what will happen in the future depends on factors such as the level of immigration and the success of the process of integration. For antisemitism, it is primarily events in the Middle East conflict that may have an impact on the attitudes of Norwegians.

Bibliography

ADL-survey (2004). Attitudes toward Jews, Israel and the Palestinian-Israeli Conflict in Ten European Countries. New York.

Anti-Defamation League (2014). Global 100, Index of 11 antisemitic statements.

Bachner, H., Ring J. (2005). Antisemitic Images and Attitudes in Sweden (English Summary). Stockholm: Forum för levande historia. https://www.levandehistoria.se/sites/default/files/material_file/antisemitiska-attityder-rapport.pdf

Bergmann, W., Erb, R. (1997, 1991). Anti-Semitism in Germany. The Post-Nazi Epoch since 1945. New Brunswick: Transaction Publ.; German edition: Der Antisemitismus in der Bundesrepublik Deutschland von 1945–1989. Ergebnisse der empirischen Forschung. Opladen: Leske+Budrich.

Bogardus, E. S. (1947). Measurement of Personal-Group Relations. Sociometry, 1(4), 306–311.

Breckler, S. (1984). Empirical validation of affect, behavior and cognition as distinct components of attitude. Journal of Personality and Social Psychology, 47(6), 1191–1205.

Couch, A. Keniston, K. (1960). Yeasayers and naysayers. Agreeing response set as a personality variable. Journal of Abnormal and Social Psychology, 60(2), 151–174.

Glenn, N. D. (1977). Cohort Analysis. Quantitative Applications in the Social Sciences, vol. 5. Newbury Park: Sage.

Heitmeyer, W. (2003). Gruppenbezognene Menschenfeindlichkeit. Die theoretische Konzeption und empirische Ergebnisse aus 2002 sowie 2003. In Heitmeyer, W. (Ed.). Deutsche Zustände, Folge 2 (p. 13–31). Frankfurt/M: Suhrkamp.

Hellevik, O. (2009). Linear versus logistic regression when the dependent variable is a dichotomy. Quality & Quantity, 43(1), 59–74.

Hellevik, O, Hellevik, T. (2017). Utviklingen i synet på innvandrere og innvandring i Norge. Tidsskrift for Samfunnsforskning, 58(3), 250–283. https://www.idunn.no/tfs/2017/03/utviklingen_i_synet_paa_innvandrere_og_innvandring_i_norge

Heyder, A., Iser, J., Schmidt, P. (2005). Israelkritik oder Antisemitismus? Meinungsbildung zwischen Öffentlichkeit, Medien und Tabus. In Heitmeyer, W. (Ed.), Deutsche Zustände, Folge 3 (144–165). Frankfurt/M: Suhrkamp.

Hoffmann, C., Kopperud, Ø., Moe, V. (Eds.) (2012). Antisemittisme i Norge? Den norske befolkningens holdninger til jøder og andre minoriteter. Oslo: Center for the Study of Holocaust and Religious Minorities.

Hoffmann, C., Moe, V. (Eds.). (2017). Attitudes towards Jews and Muslims in Norway 2017. Oslo: Center for the Study of Holocaust and Religious Minorities. https://www.hlsenteret.no/english/research/jewish-history-and-antisemitism/Population%20survey%3A%20Attitudes%20towards%20Jews%20and%20Other%20Minorities/index.html

Imhoff, R., Recker, J. (2012). Differentiating Islamophobia: Introducing a New Scale to Measure Islamoprejudice and Secular Islam Critique. Political Psychology, 33(6), 811–824.

Inglehart, R. (1977). The Silent Revolution—Changing Values and Political Styles Among Western Publics. Princeton: Princeton University Press.

Kaplan, E. H., Small, C. (2006). Anti-Israel sentiment predicts anti-Semitism in Europe. Journal of Conflict Resolution, 50(4), 548–561.

Kempf, W. (2015). Israelkritik zwischen Antisemitismus und Menschenrechtsidee. Eine Spurensuche. Berlin: Verlag Irena Regener.

Mood, C. (2010). Logistic regression: Why we cannot do what we think we can do, and what we can do about it. European Sociological Review, 26(1), 67–82.

Schuman, H., Presser, S. (1996). Questions and Answers in Attitude Surveys. Experiments on Question Form, Wording, and Context. London: Sage.

Staetsky, L. D. (2017). Antisemitism in contemporary Great Britain. London: Institute for Jewish Policy Research.

Zick, A., Küpper, B., Hövermann, A. (2011). Intolerance, Prejudice and Discrimination. A European Report. Berlin: Friedrich Ebert Stiftung.

Zick, A., Wolf, C., Küpper, B., Heitmeyer, W., Davidow, E., Schmidt, P. (2008). The Syndrome of Group-Focused Enmity: The Interrelation of Prejudices Testes with Multiple Cross-Sectional and Panel Data. Journal of Social Issues, 62(2), 363–383.

Appendix: N and test of robustness

Table A1:

N for Figures 4.5, 4.11, 4.13, 4.14 and Table 4.10 (Population sample 2017)

Combined index forScale0123
LowHigh
Antisemitism0–313421366136
Low-High147897
Islamophobia0–3885234217239
Low-High1119 456
Table A2:

N for Table 4.21 and 4.22 (Populations samples 2011 and 2017 combined)

XenophobiaWhich party supported in the conflictSum
Solely /Mostly IsraelTo some extent IsraelBoth/ noneTo some extent PalestiniansSolely / Mostly Palestinians
Very high68222463159426
High34221643150301
Medium583227170105536
Low5632386104174752
Very low5832544136312 1082
Sum27414016113727003097
Table A3.

Multivariate regression analysis with different versions of the antisemitism and Islamophobia indices as dependent variable (index with values 0–1, 0–3 or 0–3 versions narrowly and broadly defined. Regression coefficients for the last three divided by 3. Population sample 2017)*

VariableHigh value (index scores)Antisemitism indexIslamophobia index
0–10–3NarrowBroad0–10–3NarrowBroad
GenderFemale-0.044-0.052-0.024-0.090-0.051-0.036(-0.013)-0.064
Age-44 years(-0.005)(-0.017)(0.001)-0.026-0.051-0.064-0.029-0.073
EducationUniversity(-0.017)(-0.019)(-0.007)-0.026(-0.008)(-0.022)-0.026-0.033
Belief in GodYes(0.019)0.021(0.002)(0.012)-0.026(0.012)(-0.001)(0.012)
Middle East conflict (2 dummy variables) (Reference group: No opinion)Support Israel(-0.013)-0.032-0.018(-0.017)0.1390.1440.0600.125
Support Palestinians0.0540.0580.0170.079-0.080-0.085(-0.009)-0.109
Pro-Israeli attitudesStrong (5–8)(-0.019)(-0.022)(0.005)(-0.028)(0.006)(-0.012)(0.020)(-0.004)
Anti-Israeli attitudesStrong (5–8) 0.103 0.104 0.032 0.132 0.0550.0680.0320.208
Pro-Palestinian attitudesStrong (5–8)(-0.016)(-0.017)(-0.009)(0.007)(-0.003)(-0.008)-0.025(0.009)
XenophobiaStrong (7–12) 0.146 0.136 0.057 0.155 0.399 0.341 0.262 0.305
Scepticism towards immigrantsStrong (3–4)(0.010)0.024(0.011)0.039 0.364 0.347 0.163 0.303
Explained variance (adjusted Rsq)0.1300.1980.0960.2110.4190.5130.4090.450

* Regression coefficients in brackets: Not significant (5% level). In bold: Two strongest effects. Distribution on the indices: see Figure 4.11 (antisemitism) and 4.12 (Islamophobia). Meaning of broad and narrow definition: see section 3.5.

1The questionnaire in Norwegian is an appendix in the report from the study: Christhard Hoffmann and Vibeke Moe, eds., Attitudes towards Jews and Muslims in Norway 2017 (Oslo: Center for Studies of the Holocaust and Religious Minorities, 2017): 126–153, https://www.hlsenteret.no/english/research/jewish-history-and-antisemitism/Population%20survey%3A%20Attitudes%20towards%20Jews%20and%20Other%20Minorities/index.html
2More information on samples and response rates is given in the report. Hoffmann and Moe, eds., Attitudes: 22–25.
3Most of the tables and figures in this chapter show distributions for the entire samples (N=1,522 for 2011 and 1,575 for 2017). For tables/figures containing distributions for subgroup where Ns are not included, they are given in Table A1 and A2 in the appendix.
4See Werner Bergmann, “How do Jews and Muslims in Norway perceive each other? Between prejudice and cooperation”, in the present volume.
5Howard Schuman and Stanley Presser, Questions and Answers in Attitude Surveys. Experiments on Question Form, Wording, and Context (London: Sage, 1996).
6Steven Breckler, “Empirical validation of affect, behavior and cognition as distinct components of attitude”, Journal of Personality and Social Psychology 47, no. 6 (1984): 1191–1205.
7These response categories were used instead of agree-disagree scales in order to reduce the amount of yea-saying (response acquiescence), as discussed among others by Arthur Couch and Kenneth Keniston, “Yeasayers and naysayers. Agreeing response set as a personality variable”, Journal of Abnormal and Social Psychology 60, no. 2 (1960): 151–174.
8The sum for scores 1 and 2 in 2011 is 9.9 in Figure 4.1. The percentage 9.8 reported for a high score is the result when two decimals are used in the calculations, as is done in this and other figures/tables.
9Emory S. Bogardus, “Measurement of Personal-Group Relations”, Sociometry 10, no. 4 (1947): 306–311. Two of his items were the following: Accept a person “in my close circle of friends”, “as neighbours in the same street”.
10See Anti-Defamation League, Global 100, Index of 11 antisemitic statements, 2014; Werner Bergmann and Rainer Erb, Anti-Semitism in Germany. The Post-Nazi Epoch since 1945 (New Brunswick: Transaction Publ., 1997), German edition: Der Antisemitismus in der Bundesrepublik Deutschland von 1945–1989. Ergebnisse der empirischen Forschung (Opladen: Leske+Budrich, 1991); Andreas Zick et al., “The Syndrome of Group-Focused Enmity: The Interrelation of Prejudices Testes with Multiple Cross-Sectional and Panel Data”, Journal of Social Issues 62, no. 2 (2008): 363–383; Henrik Bachner and Jonas Ring, Antisemitic Images and Attitudes in Sweden (English Summary), (Stockholm: Forum för levande historia, 2005), https://www.levandehistoria.se/sites/default/files/material_file/antisemitiska-attityder-rapport.pdf . Some of the questions have been used directly, others have been partly modified.
11The difference will be less if the decrease is estimated in terms of relative rather than absolute differences (percentage points). Relative to the initial value, the decrease from 20.2 to 13.3 represents a 34% reduction, while the decrease from 7.8 to 5.5 in relative terms is 29%.
12Christhard Hoffmann, Øivind Kopperud and Vibeke Moe, eds., Antisemitism in Norway? The Attitudes of the Norwegian Population Towards Jews and Other Minorities (Oslo: Center for Studies of the Holocaust and Religious Minorities, 2012), 54–56.
13For a discussion of the arbitrariness in defining cutting points and the validity of the antisemitism construct, see Bergmann and Erb, Anti-Semitism in Germany. The Post-Nazi Epoch since 1945, Appendix 1: Problems in the Development of Anti-Semitism Scales (1997): 326–337.
14For the prejudice index, the score 11, in addition to score 12, is counted as high in the narrow definition. Score 11 means that five of the six statements are seen as completely fitting and the remaining one as somewhat fitting with one’s own opinion.
15To see whether the results of our analyses depend on the how the sub-indices are dichotomised, we have made robustness tests using the three alternative indices of antisemitism from Figure 4.6 as well as the dichotomised version as dependent variables in multivariate regression analyses, with quite similar patterns for the effects of a set of independent variables (Table A3 in appendix).
16Ottar Hellevik and Tale Hellevik, “Utviklingen i synet på innvandrere og innvandring i Norge” (“Changes in the opinion on immigrants and immigration in Norway”), Tidsskrift for Samfunnsforskning 58, no. 3 (2017): 250–283, https://www.idunn.no/tfs/2017/03/utviklingen_i_synet_paa_innvandrere_og_innvandring_i_norge.
17In contrast to our situation when selecting statements for the index on prejudice against Jews, there are few international attempts to construct such indices with regard to prejudice against Muslims. One example is the index presented in the article by Ronald Imhoff and Julia Recker, “Differentiating Islamophobia: Introducing a New Scale to Measure Islamoprejudice and Secular Islam Critique”, Political Psychology. 33, no. 6 (2012): 811–824.
18As for antisemitism, the effects of the various definitions of a high score on the sub-indices are tested in a multivariate regression analysis. For Islamophobia, the results are also quite similar (Table A3 in appendix).
19The Group-Focused Enmity Studies find stronger correlations between antisemitism and Islamophobia in 2003: r=.58 for the West Germans, r=.60 for the East Germans. Wilhelm Heitmeyer, “Gruppenbezognene Menschenfeindlichkeit. Die theoretische Konzeption und empirische Ergebnisse aus 2002 sowie 2003”, in Wilhelm Heitmeyer, ed., Deutsche Zustände, Folge 2, Frankfurt/M: Suhrkamp (2003): 19.
20Another illustration of how antisemitism and Islamophobia tend to go together is found when we look at the attitudes of the voters of Norwegian parties. There is a clear tendency that the higher the level of antisemitism in a group of voters, which in the population sample from 2017 varied between 0.7% and 13.3% for a high value on the combined index, the higher the level of Islamophobia, varying between 4.5% and 63.2%. See Hoffmann and Moe, eds., Attitudes: 99–100.
21The statement on the right to a state of their own is seen here as pro-Palestinian in its content, since it is for the Palestinians that such a right is not fulfilled at present. Since there are groups that do not accept Israel’s right to existence, the statement may also be seen as pro-Israel. This is reflected by the positive responses in the Jewish sample. The interpretation of the statement as pro-Palestinian is, however, supported by the results of the factor analysis.
22Researchers have so far used various items and indices to measure hostility towards Israel without yet reaching a consensus. Edgar H. Kaplan and Charles Small, “Anti-Israel sentiment predicts anti-Semitism in Europe”, Journal of Conflict Resolution 50, no. 4 (2006): 548–561, used the data of the ADL survey, “Attitudes toward Jews, Israel and the Palestinian-Israeli Conflict in Ten European Countries”, New York 2004; for Germany see the study of Aribert Heyder, Julia Iser and Peter Schmidt, “Israelkritik oder Antisemitismus? Meinungsbildung zwischen Öffentlichkeit, Medien und Tabus”, in Wilhelm Heitmeyer, ed., Deutsche Zustände, Folge 3 (Frankfurt/M. 2005): 144–165; L. Daniel Staetsky, Antisemitism in contemporary Great Britain (London: Institute for Jewish Policy Research, 2017); Wilhelm Kempf, Israelkritik zwischen Antisemitismus und Menschenrechtsidee. Eine Spurensuche (Berlin 2015). The items used in our study were first used by Werner Bergmann and Rainer Erb for the construction of an anti-Zionism index: Werner Bergmann and Rainer Erb, Antisemitismus in der Bundesrepublik Deutschland. Eretgebnisse der empirischen Forschung von 1946 bis 1989, Opladen 991 (English edition: Anti-Semitism in Germany. The Post-Nazi Epoch since 1945, Chapter: “Antizionism and Antisemitism”, 182–191). (Bergmann and Erb, “Antizionism and Antisemitism”, in Anti-Semitism in Germany. The Post-Nazi Epoch since 1945, English).
23Principal component analysis with varimax rotation. A similar analysis in 2011 with four statements in addition to these six produced the same dimensional solution.
24A minor change was made to the wording of the question. In 2011, the wording used was “When you think about xx, what type of contact do you think you would feel comfortable with? To what extent would you like or dislike …?” In 2017: “We will now ask you some questions about contact with people of different nationalities and religions. To what extent would you like or dislike …?”
25The questions were copied from the Norsk Monitor surveys, which were previously used in analyses of trends in attitudes of Norwegians towards immigrants; see Hellevik and Hellevik, Utviklingen. Norsk Monitor uses telephone interviews and postal questionnaires, whereas our survey is a web survey. Nonetheless, the results are quite similar both with regard to the level and with regard to the trend in scepticism towards immigrants from foreign cultures.
26But the ambiguity of one of the indicators, the statement supporting the right to a state for both parties to the conflict, may also have contributed to this result.
27Andreas Zick, Beate Küpper and Andreas Hövermann, Intolerance, Prejudice and Discrimination. A European Report (Berlin: Friedrich Ebert Stiftung, 2011).
28For arguments for using linear instead of loglinear regression analysis with a dichotomised dependent variable, see Ottar Hellevik, “Linear versus logistic regression when the dependent variable is a dichotomy”, Quality & Quantity 43, no. 1 (2009): 59–74, and Carina Mood, “Logistic regression: Why we cannot do what we think we can do, and what we can do about it”, European Sociological Review 26, no. 1 (2010): 67–82.
29With one decimal 6.0 / 2.2 = 2.7 for antisemitism and 34.2 / 10.8 = 3.2 for Islamophobia.
30Ronald Inglehart, The Silent Revolution—Changing Values and Political Styles Among Western Publics (Princeton: Princeton University Press, 1977).
31Norman D. Glenn, Cohort Analysis. Quantitative Applications in the Social Sciences, vol. 5. (Newbury Park: Sage, 1977).
32Since this is a time series and not a panel study, the respondents are not the same in 2011 and 2017. The results thus are estimates of the net changes taking place within a cohort.

Idunn bruker informasjonskapsler (cookies). Ved å fortsette å bruke nettsiden godtar du dette. Klikk her for mer informasjon